WellDataBaseML example: Predicting Well production within Eagle Ford Play

A problem demonstrating how WellDataBaseML can be applied to predict oil/gas production.

The machine-lerning problem described here is related to the clasical history matching problems performed regularly for oil/gas exploration.

If WellDataBaseML is not installed, first execute import Pkg; Pkg.add("WellDataBaseML"); Pkg.add("WellDataBase").

Import WellDataBaseML

Load Eagle Ford Play datasets

The loaded data are provided in the follwoing form:

Machine learning (ML) analysis

Machine learning (ML) analysis can be performed to predict well production based on this dataset.

Under this example, the goal is to use all the data collected up to 2015 for training.

After that, the applied ML methodology performs blind prediction for 1 year ahead (2015-2016).

Setup training prediction / prediction dates

Execute ML run

ML analyses is excuted as follows:

The figures and tables above represent training and prediction summaries.